Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add instructions/dpup/meshcore-mcp/agents-mdgit clone --depth 1 https://github.com/dpup/meshcore-mcpWhat it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.02794 | $0.02794 |
| Opus 5 | $0.01397 | $0.01397 |
| Sonnet 5 | $0.00559 | $0.00559 |
| Haiku 4.5 | $0.00279 | $0.00279 |
Grade A, and why
meshcore-mcp AGENTS.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 196 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
Guidance for AI agents (and humans) working on meshcore-mcp. Keep this current when architecture or conventions change.
What this is
A Model Context Protocol (MCP) server that exposes a MeshCore node — and the
mesh reachable through it — as a clean, high-signal surface of tools,
resources, and prompts for any MCP client (Claude Code, or meshcore-elmer's
bridge). It is the device layer, and ungated by design: it contains no
conversation policy, admin-channel gate, or autonomous behavior (that belongs to
meshcore-elmer). See docs/plans/ for the PRD (v0.3) and execution plan.
Docs are three layers: README.md (landing + the Claude Code MCP-config snippet
- the tool/resource/prompt tables),
docs/guide.md(hand-written concepts & recipes), anddocs/api.md(generated by TypeDoc);llms.txtindexes all three for agents.
Companion projects, by the same author — keep all three consistent in stack, style, and tone:
@dpup/meshcore-ts— the typed device client this server is built on (runtime dependency).@dpup/meshcore-sim— the deterministic simulator we test and demo against (dev dependency; never shipped).
The two contracts (read this first)
meshcore-mcp is a thin server wedged between two contracts it does not own.
Below — @dpup/meshcore-ts. The server holds one persistent MeshCoreClient
(autoSync: true) and consumes its normalized surface: named events
(contactMessage, channelMessage, channelData, advert, …), typed models,
hex-string keys, Date timestamps, typed errors. It never touches
@liamcottle/meshcore.js directly. MeshService takes its MeshCoreClient by
injection — production builds MeshCoreClient.tcp(host, port); tests build
new MeshCoreClient(sim.asConnection()). Nothing below MeshService knows which.
Above — MCP, via @modelcontextprotocol/sdk's McpServer: tools
(registerTool, Zod schemas, annotations), resources (registerResource, the
subscribable live stream), prompts (registerPrompt), over a
StdioServerTransport.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 2d ago First seen · 196 lines · 2,794 tokens per session scan A cfbd3af3ae63
meshcore-mcp AGENTS.md is an instructions file published in the GitHub repository dpup/meshcore-mcp (1 stars, last pushed 3mo ago), licensed MIT. It adds 2,794 tokens to every session, about $0.0140 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other instructions, from other repositories
rosclaw AGENTS.md
Instructions for ros-claw/rosclaw, covering rosclaw agent instructions, runtime boundary, tool policy, robot integration setup and capability apps.
rosclaw CLAUDE.md
Instructions for ros-claw/rosclaw, covering claude.md — rosclaw project onboarding, rosclaw runtime boundary (managed), safety contract (p0), robot integration setup and capability apps.
open-xiaoai-bridge AGENTS.md
Instructions for coderzc/open-xiaoai-bridge, covering agents.md, 系统架构, 项目结构, 核心组件 and mainapp (core/app.py).
copperhead AGENTS.md
Instructions for copperheadhq/copperhead, covering agents.md, repository overview, sources of truth, build and verification and architecture.
copperhead CLAUDE.md
Instructions for copperheadhq/copperhead, covering claude.md, what this repo is, sources of truth, workflow (openspec) and architecture (per spec.md §2).
chiplab AGENTS.md
Instructions for veecle/chiplab, covering agents.md, what is chiplab, 1. install / connect chiplab, 2. verify the connection and 3. pick the example for the board you want.